Figure 4: Optimal decision boundaries computed numerically by DP. | Nature Communications

Figure 4: Optimal decision boundaries computed numerically by DP.

From: Optimal policy for value-based decision-making

Figure 4

(a) The decision boundaries and choices at three representative time points (t=0.5, 1 and 2 s), shown in the space of reward estimates . (b) The same as a, but shown in a rotated space, . (c) The decision boundaries in terms of the expected reward difference, , as functions of time. The distance between boundaries decreases as time elapses, and collapses to zero at some point. (d) The same decision boundaries shown in the space of accumulated evidence, which is represented by the particle location in a DDM. The cyan trace is a sample particle trajectory, representing evidence accumulation and the subsequent choice of option 1.

Back to article page